How to find the closest common parent?

Neo4j Graph Platform Cypher
1

Hello,
I assume my problem is not a new one, but I can't find an elegant solution.
merge (A:LEU {LEID: "A"})
merge (B:LEU {LEID: "B"})
merge (c:LEU {LEID: "C"})
merge (aa:LEU {LEID: "AA"})

merge (a)-[:CONTROLS {share:toFloat(100.0)}]->(b)
merge (a)-[:CONTROLS {share:toFloat(100.0)}]->(c)
merge (aa)-[:CONTROLS {share:toFloat(100.0)}]->(a)

I run the following code, it works.

MATCH (n:LEU)

WHERE (:LEU{LEID:"B"})<-[:CONTROLS0..]-(n) and (:LEU{LEID:"C"})<-[:CONTROLS0..]-(n)

RETURN collect(n)[0] as nn

My demand is about how to create dynamically the WHERE condition. I try the following, but it doesn't work. I assume that I must mix UNWIND, WHERE or FOREACH but how ?
MATCH (n:LEU)
WHERE EXISTS((n:LEU)<-[:CONTROLS]-(:LEU)) and NOT EXISTS((:LEU)<-[:CONTROLS]-(n:LEU))
WITH collect(n) as children
UNWIND children as child
MATCH (parent:LEU)
WHERE (child)<-[:CONTROLS*0..]-(parent)
RETURN parent

2

Hi @depire,

Please explain what do you mean by dynamic Where ..
If you tells us exactly what the requirement is then group can help you much better way

3

Here is your starting scenario:

Screen Shot 2020-04-15 at 11.56.39 PM

Query:
MATCH (b:LEU)-[:CONTROLS]->(n:LEU)
with collect(n) as nm
unwind nm as nm1
RETURN nm1

Result:

Screen Shot 2020-04-15 at 11.57.56 PM

Query:

MATCH (b:LEU)-[:CONTROLS]->(n:LEU)
WHERE NOT EXISTS ((n)-[:CONTROLS]->(:LEU))
with collect(n) as nm
unwind nm as nm1
RETURN nm1

Result:

Screen Shot 2020-04-15 at 11.58.34 PM

Query:

MATCH (n:LEU)
WHERE EXISTS ((:LEU)-[:CONTROLS]->(n))
AND NOT EXISTS ((n)-[:CONTROLS]->(:LEU))
WITH collect(n) as children
UNWIND children as child
MATCH (parent:LEU)
WHERE (child)<-[:CONTROLS*0..]-(parent)
RETURN parent

Result:

Screen Shot 2020-04-15 at 11.59.19 PM
Screen Shot 2020-04-15 at 11.59.19 PM726×486 21.7 KB

Hope this helps.

4

Here is my solution, one can takes directly the first common parent.
MATCH (n:LEU)
WHERE EXISTS((n:LEU)<-[:CONTROLS]-(:LEU)) and NOT EXISTS((:LEU)<-[:CONTROLS]-(n:LEU))
WITH collect(n) as children
MATCH (first_common_parent:LEU)
WHERE ALL(child in children WHERE ((child)<-[:CONTROLS*0..]-(first_common_parent)))
RETURN collect(first_common_parent)[0]

"Dynamically" means that the user mustn't enter manually the children.

That may be could be useful for somelse.

5

That's really great that you solved the problem by yourself :metal:.
Please mark the post as solution so that others could use it

6 
match (a:LEU)-[r:CONTROLS]-(b:LEU)
with a, count(r) as cnt where cnt > 1
return a

This returns node "A"